通信与信息技术Issue(3)：134-140,7.
基于PCA和改进的FCNN的CSI室内定位算法
CSI localization algorithm based on PCA and improved FCNN
摘要
Abstract
In the field of indoor positioning,satellite positioning system cannot be accurately applied because buildings block satel-lite signals.To solve this problem,an channel state information(CSI)indoor localization method based principal component analysis(PCA)and improved fully convolutional neural network(FCNN)model is designed.Firstly,the CSI data of the location region is collected,and the filtering algorithm is used to pre-process the data.After removing the noise in the signal,the PCA is used to extract the feature of the CSI data to reduce the feature dimension and improve the performance of the model.Secondly,a particle swarm optimization(PSO)al-gorithm is improved to enhance the ability of particle optimization.Finally,the improved PSO is combined with error back propagation(BP)as a training algorithm for the FCNN model,and the location information is trained and output to realize CSI indoor positioning.The experimental results show that the average positioning error of the proposed algorithm is about 0.034 m,which is improved by 85.71%,21%and 46.3%compared with the original FCNN without PCA,the original FCNN with PCA,and time convolution neural network(TCNN)with PCA,respectively.关键词
信道状态信息/室内定位/粒子群算法/主成分分析/误差反向传播Key words
Channel state information/Indoor localization/Particle swarm algorithm/Principal component analysis/Error back prop-agation分类
信息技术与安全科学引用本文复制引用
覃圣,罗珊珊,裴氏莺..基于PCA和改进的FCNN的CSI室内定位算法[J].通信与信息技术,2025,(3):134-140,7.基金项目
广西壮族自治区科技计划项目(桂科:AD21238038) （桂科:AD21238038）
